[Pacemaker] seemingly erratic stonith behavior

[David Vossel]

----- Original Message -----
> From: "David Pendell" <lostogre at gmail.com>
> To: "The Pacemaker cluster resource manager" <pacemaker at oss.clusterlabs.org>
> Sent: Monday, December 31, 2012 6:09:37 PM
> Subject: [Pacemaker] seemingly erratic stonith behavior
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 
> I have a pacemaker cluster setup with two fencing levels. ipmi for
> the first and an apc pdu for the second. I am using the
> fencing-level directive in the cib and that works. Unfortunately, I
> am having a problem with how the fencing plays out when i take the
> ipmi system offline. it goes like this...
> 
> ipmi... fail
> ipmi... fail
> ipmi... fail
> pdu.... success
> ipmi... fail
> ipmi... fail
> 
> So even though I am getting success on the pdu, it continues to try
> to fence with ipmi.

We can't help explain this behavior with out a crm_report, otherwise we are just guessing why this occurs.  I agree that I would not expect ipmi to keep getting tried after pdu success.

-- Vossel

> I wouldn't consider this a problem, except that
> when the ipmi fence tries again after the pdu succeeds, the ipmi
> isn't ready and it hard locks the machine. I am going to talk to the
> manufacturer on Wednesday, but this seems like a bug to me.... I
> would expect that once a fencing method succeeds, it would quit
> trying.
> 
> d.p.
> 
> _______________________________________________
> Pacemaker mailing list: Pacemaker at oss.clusterlabs.org
> http://oss.clusterlabs.org/mailman/listinfo/pacemaker
> 
> Project Home: http://www.clusterlabs.org
> Getting started:
> http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf
> Bugs: http://bugs.clusterlabs.org
>